Advanced search

Message boards : News : GPUs supported by the "ACEMD Long" application

Author Message
Profile MJH
Project administrator
Project developer
Project scientist
Send message
Joined: 12 Nov 07
Posts: 696
Credit: 27,266,655
RAC: 0
Level
Val
Scientific publications
watwat
Message 28976 - Posted: 4 Mar 2013 | 18:49:48 UTC

Starting with this update to the ACEMD Long application, we are no longer supporting GPUs of compute capability 1.2 and lower. In practice this is everything below the spec of the GTX260 (8000-series, 9000-series, and low 2x0 and 3x0 cards).

There are two reasons why we are doing this:

1) It simplifies the application development because we have to support fewer hardware generations.
2) We are able to improve the specification of the Long WUs to better reflect the majority of the GPUs crunching them.

We are very grateful for everyone's contribution to GPUGRID, and I hope none of you with one of these older cards is upset by the exclusion. Work will still be available for you on the short queues.

Thanks all,

MJH

tomba
Send message
Joined: 21 Feb 09
Posts: 497
Credit: 700,690,702
RAC: 0
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 28983 - Posted: 4 Mar 2013 | 20:41:21 UTC - in response to Message 28976.

I finished a TONI about 30 minutes ago. Since then I've not had a new WU despite many "Update"s. Does this mean my GTX 460 is past its prime; no more "Long"s??
____________

flashawk
Send message
Joined: 18 Jun 12
Posts: 241
Credit: 1,694,210,922
RAC: 703,764
Level
His
Scientific publications
watwatwatwatwatwatwatwat
Message 28985 - Posted: 4 Mar 2013 | 21:19:00 UTC - in response to Message 28983.

He clearly says 200 series and below, nobody getting any work. I have 3 GPU's sitting idle right now, it's all over the forum.

ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2671
Credit: 753,325,024
RAC: 489,512
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 28992 - Posted: 5 Mar 2013 | 21:29:52 UTC - in response to Message 28985.

I have 3 GPU's sitting idle right now

OT: I'd set up a backup project in this case.

MrS
____________
Scanning for our furry friends since Jan 2002

Wiyosaya
Send message
Joined: 22 Nov 09
Posts: 111
Credit: 172,459,353
RAC: 376,862
Level
Ile
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29089 - Posted: 9 Mar 2013 | 3:52:21 UTC - in response to Message 28983.

I finished a TONI about 30 minutes ago. Since then I've not had a new WU despite many "Update"s. Does this mean my GTX 460 is past its prime; no more "Long"s??

My 460 has been getting long queue tasks since the change to the new app. Are you sure your driver supports the latest long queue app? I'm on the 310.70 driver.
____________

flashawk
Send message
Joined: 18 Jun 12
Posts: 241
Credit: 1,694,210,922
RAC: 703,764
Level
His
Scientific publications
watwatwatwatwatwatwatwat
Message 29091 - Posted: 9 Mar 2013 | 5:42:09 UTC - in response to Message 28992.

I have 3 GPU's sitting idle right now

OT: I'd set up a backup project in this case.

MrS


I've done that in the past and got inundated with so many tasks from the other project that it took me a week to clear them out. I thought I had my preferences set correctly for that project but it was like opening a flood gate, I made the changes in my account for the least amount of work and then updated BOINC on my computers. Needless to say, I was peeved, I wasn't going to abort all the work units when GPUGRID had more work. I'm sure I must have done something wrong. It gets flat cold in my house when those GPU's go off line, it's been getting in to the low twenty's, upper teens F here at night at 5000 feet in the Sierra's.

ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2671
Credit: 753,325,024
RAC: 489,512
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29105 - Posted: 10 Mar 2013 | 13:22:43 UTC - in response to Message 29091.

I've been using it with BOINC 7.0.25, 7.0.28 and 7.0.44 and it works well: just set the ressource share of the backup project to 0 and if there's no other work BOINc will (almost) reliably fetch single WUs from that other project. For Milkyway I'm using share 1 to get a cache of 40 WUs, but each of them takes <1 min.

MrS
____________
Scanning for our furry friends since Jan 2002

Post to thread

Message boards : News : GPUs supported by the "ACEMD Long" application